ChangeLog

=== mptopcom 1.2 ===

-- general --
 * New binary random_walk for doing a random walk on the graph of all
   trinagulations.
 * New binary canonicalRepresentative for finding the canonical representative
   of an input triangulation.
 * New parameter --lex-min for switching from lex maximal representatives to
	lex minimal.
 * Fix error when computing full triangulations modulo symmetry, resort to
   brute force for now.
 * polymake > 4.1 is now strictly required
 * Using standard mptopcom with more than 64 points will now give an error
 * For more than 64 points, build mptopcom with -DWITH_PM_BITSET
 * -DWITH_PM_BITSET only works with polymake > 4.2
 * --print-secondary-cone print the rays of the secondary cone after every
	triangulation, does not include lineality space

-- src --
 * Testsuite refactored for easier debugging and flexibility.
 * Fix bug causing memory to leak when --regular was enabled
 * Move several code pieces to polymake, as they are used in other algorithms
	and we want to avoid duplication


=== mptopcom 1.1 ===

-- general --
 * The output for the symmetric case was changed, it is now
   (canonical triangulation, canonical gkz vector).
 * Introduce --central
   - Computes only those triangulations with maximal first gkz coordinate.
	- Depends on ordering of points
 * Introduce --cyclic-flips
   - Works only for cyclic polytopes
	- Avoids computing gkz vectors by utilizing combinatorial description of
	  flips
 * Command line options are parsed differently, unrecognized options trigger
   an error
 * If group generators are malformated, an error is printed and mptopcom or
   TOPCOM exits. Previous behaviour was to continue with no group.
 * Will build single-threaded version if no MPI present
 * Add canonicalRepresentative binary that computes a canonicalRepresentative
   for a given triangulation.
 * Add randomBFS binary that executes a randomized breadth first search on the
   flip graph of a triangulation.
 * Introduce --version to print current version
 * Option -F got renamed to --full for consistency

-- src --
 * Replace TOPCOM's IntegerSet with polymakes BitSet
   - Other polymake Set types can be used as well
 * Introduce LongBitset based on just a long
 * Separate code for cyclic polytopes triggered by --cyclic-flips
   - Use combinatorial description of up- and downflips
	- Files: lib-src-ReverseSearch/AugmentedPointConfigurationCyclic.hh
	         lib-src-ReverseSearch/FlipDataCyclic.hh
 * Refactor the code using template meta programming
   - AugmentedPointConfiguration contains the points with extra data
	  (chirotope, GKZ object, ...)
	- FlipChecker will check whether a flip should be used in the graph, e.g.
	  --regular: Only flips between regular triangulations are considered
	  --central: Only flips that preserve the first coord are considered
	- FlipData generates the flips
 * Scalar = pm::Integer is now possible
 * Group closure is now computed via polymakes permlib interface
 * Several small bugfixes

=== mptopcom 1.0 ===

-- general --
 * license is now GPLv3
 * Add README.md for github
 * Add binaries
	- mptopcom1: Single threaded reverse search
	- mptopcom: Parallelized (budgeted) reverse search

-- src --
 * add source for reverse search, GKZ vectors, switch tables, LRU cache
 * add mts for parallelizing reverse search

-- changes wrt TOPCOM 0.17.8 --
 * build system changed to ninja
 * using system gmp and cdd
   - gmp and cdd not shipped anymore
	- configure script checks for gmp
 * configure checks for polymake and ompi
 * configure script is derived from polymake's configure script, hence only
   supports the same compilers as polymake
